Though this behavior is not expected , I'll suggest not to use chef as a
file server to distribute large binaries or config files . use NFS or a
vanilla lighthttpd server in to host the file and remotely_file resource
inside the chef recipes to install it instead.
For getting the root cause you can use ruby-prof with kcachegrind to check
which part of the close is actually consuming large memory
Regarding the file server bit, we're using chef-solo+git specifically to
avoid setting up infrastructure (like chef or nfs or lighttpd servers) just
to manage our infrastructure.
I could easily workaround the bug by running cp source destination in a
bash/execute resource but it seems like the cookbook_file resource was
meant to be used for this :?
Is there a howto on using ruby-prof to find the problem here or chef
problems in general?
Though this behavior is not expected , I'll suggest not to use chef as a
file server to distribute large binaries or config files . use NFS or a
vanilla lighthttpd server in to host the file and remotely_file resource
inside the chef recipes to install it instead.
For getting the root cause you can use ruby-prof with kcachegrind to check
which part of the close is actually consuming large memory
Regarding the file server bit, we're using chef-solo+git specifically to
avoid setting up infrastructure (like chef or nfs or lighttpd servers) just
to manage our infrastructure.
I could easily workaround the bug by running cp source destination in a
bash/execute resource but it seems like the cookbook_file resource was
meant to be used for this :?
Is there a howto on using ruby-prof to find the problem here or chef
problems in general?
Though this behavior is not expected , I'll suggest not to use chef as a
file server to distribute large binaries or config files . use NFS or a
vanilla lighthttpd server in to host the file and remotely_file resource
inside the chef recipes to install it instead.
For getting the root cause you can use ruby-prof with kcachegrind to
check which part of the close is actually consuming large memory